I am a relatively young faculty member. I was asked to provide a reference for a student of mine so he is admitted in a PhD program in Sciences in UK. The student is probably one of the best students I had in a couple of years in terms of academic potential. I firmly believe he will do well as he is bright and hard-working.

I have to fill in an online form rating him. My problem is that the form's scaling is very strong. How can I honestly say that someone was outstanding (3%) and not exceptional (1%)? Even the third most positive option is unusually good (rated at 10%) which I feel as a characterization in itself should be strong enough for someone to be at least considered for a PhD.

My classes are never above 40 to 45 students and I believe that simply "lumping" all the students from my past couple of years in one sample is ineffective and unreasonable. I would probably be unable to single out one of my students as the single best student I've ever had so that I can place him in the top 1% interval with certainty and not top 3%.

As I said, I want to support him and I worry that simply putting him down as outstanding would harm him as other referees might be tempted to "max out" everything. On the other hand simply rating him as exceptional in almost all aspects would be just grade-inflation.

Could someone provide me with some advice on the matter and what they would do? The scaling I have come across before topped at 5% which I found reasonable but here this not the case.

  • 1% and 3% are probably more difficult to perceive than thinking if the student is about 3 (about .2%) or 2 (about 2.5%) standard deviations above the mean in a normally distributed situation. If he is really way above everyone by a lot, then I don't see why not say exceptional. I'd also ask the student for a copy of CV and examine if he has anything such as being on dean's list, scholarship, and awards to go with my recommendation; I wouldn't want my recommendation to appear overly optimistic or biased. – Penguin_Knight Dec 5 '13 at 18:54
  • @Penguin_Knight: I agree completely. After all if a recommendation letter appears biased it will work against the person it recommends. – user9805 Dec 5 '13 at 19:55

The question of gauging inflation in these situations is both serious and important. It affects peoples' lives!

The difficult task of the letter writer is to visualize what other letter-writers will say, their degree of effusiveness. "Being brutally honest" is obviously stupid and naive to the point of destructiveness.

The operational issue is about achieving the desired effect. And, in particular, if your recommendee is fairly exceptional, it would be a grievous error to praise them "modestly". Yes, the problem is to discriminate between "fairly exceptional" and "exceptional" and "truly exceptional". And these are not things that can be rationally discussed about young people by committees...

Thus, consider the noise in the system into which you put your information...

| improve this answer | |
  • 4
    This sort of system needs to change. Exceptional (1%) should mean just that, not relative to what you think a random set of other people will determine. Unfortunately this is easy to say and difficult to execute. I just hate these types of metrics that seemingly lose its value based on an ever-shifting, subjective scale. Can't we just have a metric that says 1.) Yes, I recommend this person for the program or 2.) No, I do not recommend this person for the program? – LordStryker Dec 5 '13 at 18:17
  • 5
    @LordStryker, I agree, this inflation/corruption is distressing. One aspect of the problem is the usual human-nature aspect that some people will hype their own proteges beyond reason, by comparison to which a sane, modest recommendation sounds pathetic and weak. Unfortunately, to "protest-vote" by behaving "modestly/rationally" in such a context dis-serves the students... – paul garrett Dec 5 '13 at 18:43
  • 1
    Thank you for your answer Paul. It will take it into consideration. – user9805 Dec 5 '13 at 19:51
  • 1
    Further to the above comments, you could consider the ranking scale presented to you as effectively a binary yes/no question, given you can be fairly sure other letter writers will rank their applicants in the "exceptional" category, who will most likely reason in the same way. This, of course, makes a nonsense of such an extreme scale. You then write a letter which sets out your recommendation according to your actual perceptions of the student's ability - one which you can stand by as a professional, and not be embarrassed by either over-praising, or faintly-praising. – Nicholas Dec 5 '13 at 20:46

Suppose I am in your position,

I have a student and I have to rate him. The rating in my mind is top 2%. I need to make a choice between top 1% and top 3% when I fill in the rating form.

I would choose top 1% for the following reasons,

There is really no significant difference between 1% and 2% or between 2% and 3%. I only know his past performance well. I can only firmly believe his future will be bright (I am not a superman. I cannot predict the future.). I want to give him top 1% rating because I want to let him have a better chance to succeed. I know top 1% is a little bit grade inflation. But, this is not my fault - there is no place for my real rating, that is, top 2%. I would not feel guity for the inflation. I would not know what others will do. I am just performing my duty as his professor.

| improve this answer | |

First, find out, independently of course, what percentile, 1 2 5, this "science" committee requires for admission into their hallowed PhD programme, and give them what they want. Of course your students are in the top k% -- why would be recommending them otherwise?

Second, and this is very important for the future of our planet, you need to convince the world that this "science" committee is made up of complete idiots, hypocrates, clowns, and fools who are doing all they can to help aggravate the UK's STEM crisis.

(OK, I'm being a wee bit cynical, but I provide this as a separate answer because I fear the other folks, while sympathetic to your plight, just have not come down hard enough on the ridiculously high "standards" that are being set.)

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.